Regional geochemistry and continental heat flow: implications for the origin of the South Australian heat flow anomaly

Earth and Planetary Science Letters 183 (2000) 107^120 www.elsevier.com/locate/epsl

Regional geochemistry and continental heat £ow: implications for the origin of the South Australian heat £ow anomaly Narelle Neumann *, Mike Sandiford 1 , John Foden Department of Geology and Geophysics, Adelaide University, Adelaide, S.A. 5005, Australia Received 15 May 2000; received in revised form 5 September 2000; accepted 19 September 2000

Abstract Existing measurements from South Australia define a broad ( s 250 km wide) zone of anomalously high surface heat flow (92 þ 10 mW m32 ). This zone is centred on the western margin of the Adelaide Fold Belt (Neoproterozoic to early Phanerozoic cover floored by Palaeoproterozoic to Mesoproterozoic basement), where it borders the eastern Gawler Craton and Stuart Shelf (Palaeoproterozoic^Mesoproterozoic). To the west, in the western Gawler Craton (Archaean to Palaeoproterozoic), heat flow averages V54 mW m32 while to the east in the Willyama Inliers (Palaeoproterozoic) heat flow averages V75 mW m32 . We use a regional geochemical dataset comprising s 2500 analyses to show that the anomalous heat flow zone correlates with exceptional surface heat production values, mainly hosted in Palaeoproterozoic to Mesoproterozoic granites. The median heat production of Precambrian `basement' rocks increases from 6 3 WW m33 west of the anomalous zone to V6 WW m33 within the anomalous zone. In the highest known part of the heat flow anomaly, Mesoproterozoic gneisses and granites of the Mount Painter Province in the northern Adelaide Fold Belt yield an area-integrated mean heat production of 9.9 WW m33 . These data suggest that the anomalous heat flow reflects an unusual enrichment in U and Th in this part of the Proterozoic crust, with the total complement of these elements some 2^3 times greater than would be expected for Proterozoic crust on the basis of the global heat flow database. This extraordinary enrichment has played an important role in modulating the thermal regime of the crust in this region, and particularly its response to tectonic activity. ß 2000 Elsevier Science B.V. All rights reserved. Keywords: heat £ow; regional patterns; heat sources; South Australia; Proterozoic; granites

1. Introduction The thermal structure of continental interiors is

* Corresponding author. Tel.: +61-8-8303-4868; Fax: +61-8-8303-4347; E-mail: [email protected] 1

Present address: School of Earth Sciences, University of Melbourne, Parkville, Vic. 3010, Australia.

fundamental to their long-term tectonic and geochemical evolution. Our understanding of thermal regimes within the continental crust has been greatly in£uenced by surface heat £ow measurements. These measurements are particularly important because they yield information concerning the thermal structure of the lithosphere and constrain the vertical distribution of heat sources [1^ 3]. Further, heat £ow data provide a unique insight into the geochemistry of the crust because they constrain the depth integrated abundance of

0012-821X / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 1 2 - 8 2 1 X ( 0 0 ) 0 0 2 6 8 - 5

EPSL 5622 3-11-00

108

N. Neumann et al. / Earth and Planetary Science Letters 183 (2000) 107^120

heat-producing elements [4,5]. In the recent past, considerable e¡ort has been expended towards understanding global heat £ow averages, and the source distributions that contribute to them (e.g. [3,6^8]). While such global averages are undoubtedly important, their signi¢cance should be evaluated with regard to the following points. Firstly, the so-called global heat £ow dataset is strongly biased by measurements made in North America, Europe and southern Africa, with the heat £ow ¢eld from other continental regions virtually unknown. Secondly, given that many important geological processes are temperature dependent, the natural spatial variation in thermal parameters is of greater relevance than global averages. In this regard, regions of elevated heat £ow are fundamental to our understanding of the thermal structure of the continental crust. This paper concerns itself with a region of elevated heat £ow in South Australia. Existing heat £ow measurements in this region suggest either anomalous mantle activity or that radiogenic crustal sources contribute more than twice what would be expected on the basis of global heat £ow averages. In this region, crustal growth occurred mainly from the Palaeoproterozoic through to the early Mesoproterozoic [9]. Heat £ow measurements are often of poor quality or display broad scatter, and as with all such measurements there is a need to evaluate their plausibility. A primary purpose of this paper is to determine the validity of these heat £ow values in light of inferences about mantle thermal regimes and surface heat production parameters derived from regional geochemical and geophysical datasets. These data suggest that the anomalous heat £ow re£ects extraordinary concentrations of heat-producing elements in the crust. In Section 6, we brie£y explore the origin and implications of this exceptionally enriched crust. 2. Some preliminary remarks concerning global heat £ow Surface heat £ow is a measure of the combined heat £ow from the convective mantle, radiogenic heat from the decay of U, Th and K within the

lithosphere and transient perturbations associated with tectonic, magmatic, hydrologic and/or climatic activity. Over 10 000 global continental heat £ow measurements [7] have been used to constrain the chemical and thermal structure of the lithosphere. However, it should be stressed that 90% of these measurements are from three continents; Europe, North America and Africa. In comparison, the heat £ow ¢eld of South America, Australia and Asia is relatively poorly known, while Antarctica is virtually unknown. Note also that almost all heat £ow measurements within Africa come from either the Archaean provinces of southern Africa or the East African rift, with the heat £ow ¢eld of the great majority of the continent outside these provinces only very poorly known. An important question, therefore, concerns just how representative the existing heat £ow dataset is. Although the global heat £ow dataset suggests marked spatial and temporal variations, existing measurements have been used to constrain the crustal contribution to the surface heat £ow average. Using a reduced heat £ow value of 27 mW m32 , McLennan and Taylor [5] proposed that `the crustal radiogenic crustal component of continental heat £ow must lie within the range of 18^ 48 mW m32 , and almost certainly lies in the range 21^34 mW m32 '. This range correlates well with crustal heat £ow contributions of 20 to s 44 mW m32 calculated using geochemical and seismic data from Archaean and Proterozoic crustal sections together with average surface heat £ow values [3,10^14]. A major focus of previous heat £ow studies has been the relationship between surface heat £ow and tectonic age (Fig. 1). An inverse relationship is suggested by the available data which increase from 41 mW m32 for the Archaean, 51^54 mW m32 for Proterozoic areas, 61 mW m32 for the Late Palaeozoic and 72 mW m32 for Cenozoic to Mesozoic terranes [2,15,16]. High values within young tectonothermal terranes are interpreted to re£ect transient thermal perturbations associated with tectonic/magmatic activity, whereas variations between Archaean and Proterozoic terranes re£ect di¡erent crustal heat production distributions or mantle heat £ow contributions [15].

EPSL 5622 3-11-00

N. Neumann et al. / Earth and Planetary Science Letters 183 (2000) 107^120

Fig. 1. Global surface heat £ow averaged in groups according to the age of the last major tectonothermal event, from [15,16]. Age key: A^Archaean, EPr^Early Proterozoic, LPr^ Late Proterozoic, EPa^Early Palaeozoic, LPa^Late Palaeozoic, M^Mesozoic, C^Cenozoic. The dashed line represents the mean heat £ow value for that group and the boxes indicate the age ranges for the data and standard deviations of the means. The number of data in each group is listed below that box. Dark grey shading for Archaean, mid-grey for Proterozoic and light-grey for Phanerozoic data.

Previous investigations have also noted that there are signi¢cant variations in heat £ow according to their proximity to Archaean^Proterozoic boundaries, with values of V40 mW m32 for Proterozoic crust within 100^400 km of Archaean cratons increasing to 54^58 mW m32 in Proterozoic regions greater than 400 km from cratonic regions [8]. Although di¡erences in crustal heat production may contribute to the spatial variation in heat £ow, the observed signature has been primarily attributed to di¡erences in lithospheric thickness, with heat diverted around thick cold Archaean roots into the surrounding younger terrains [8,17]. While the global heat £ow average for continental lithosphere has been used as the basis for studies of the thermal evolution in many Precambrian terranes, heat £ow values within Proterozoic Australia are signi¢cantly higher than the global average for terranes of comparable, or younger age. On the basis of 90 measurements within Australia, Sass and Lachenbruch [18] identi¢ed three heat £ow provinces : the primarily Archaean

109

Western Shield Province (ages s 2500 Myr), the mainly Proterozoic Central Shield Province (600^ 2000 Myr) and the Palaeozoic Eastern Province (Fig. 2). While the average heat £ow of 39 mW m32 within the Western Shield Province is consistent with other shield areas of similar age, the Central Shield Province records signi¢cantly elevated heat £ow, with 32 determinations providing a mean of 83 þ 20 mW m32 . Heat £ow in the Eastern Province is also elevated, with 38 measurements producing an average of 72 mW m32 . Sass and Lachenbruch [18] proposed a reduced heat £ow contribution of 27 mW m32 throughout all provinces, with di¡erences in surface heat £ow re£ecting lateral variations in crustal heat production in the two western zones and additional thermal e¡ects of magmatic and tectonic activity in the Eastern Province. Such an interpretation requires that within the Central Shield Province, internal heat sources contribute s 50 mW m32

Fig. 2. Surface heat £ow map of Australia (after [18]) showing the main outcrops of Precambrian crystalline basement. The three provinces, W, C and E, represent adaptations of Sass and Lachenbruch's [18] Western Shield, Central Shield and Eastern Province, respectively. W is characterised by heat £ows mostly less than 50 mW m32 and represents the part of the continent formed principally in the Archaean. C, with heat £ows typically in the range 50^90 mW m32 , comprises portions of the continent where the principal continental growth occurred in the Proterozoic. E, with variable heat £ow character, represents terranes accreted to the Australian continent during the Phanerozoic. Note that the boundary between W and C occurs in the middle of the Gawler Craton in South Australia (see Figs. 3 and 4).

EPSL 5622 3-11-00

110

N. Neumann et al. / Earth and Planetary Science Letters 183 (2000) 107^120

to the surface heat £ow. This implies that a signi¢cant proportion of the Australian continent lies near, or beyond, the upper limit of the range of crustal heat production parameters suggested by the global heat £ow dataset (e.g. [5]). If this assertion can be corroborated, it would imply that the natural variation in heat £ow within continents should be revisited, with any inferences about global heat £ow averages cognisant of the potential bias introduced by the geography of the available data. South Australia has the most detailed heat £ow record within the Central Shield Province and therefore provides an ideal setting to assess the validity of the elevated heat £ow measurements. In order to evaluate the South Australian heat £ow record, we will ¢rst outline the geological framework of this region. 3. Geological setting South Australia is dominated by two Archaean^Proterozoic cratonic terranes separated by the Adelaide Fold Belt (Fig. 3). Unless otherwise stated, all geochronological data are from Drexel et al. [19] and Daly et al. [20]. The western region of South Australia comprises the Gawler Craton, a stable crystalline basement terrane composed of Archaean to Mesoproterozoic magmatic and metasedimentary rocks. The northern and western boundaries of the Gawler Craton are covered by the sedimentary Neoproterozoic O¤cer Basin and the Cenozoic Eucla Basin, respectively. The northeastern part of the Craton is covered by a thin veneer of Neoproterozoic sediments of the Stuart Shelf, and the eastern extent is de¢ned by the Torrens Hinge Zone (Fig. 3). The Gawler Craton has been divided into numerous subdomains based on tectonic, chemical and isotopic constraints [19]. Western subdomains are dominated by Archaean to Early Proterozoic crustforming events including the Sleaford and Mulgathing Complexes. In contrast, the eastern Gawler Craton is dominated by younger crustforming events, with Late Palaeoproterozoic magmatism in the Cleve, Lincoln and Moonta subdomains including the V1850 Ma bimodal Lincoln Batholith and V1700 Ma granites and minor

Fig. 3. Summary of main Precambrian elements of South Australian geology (adapted from Drexel et al. [19]). Dashed lines represent the inferred boundaries of the Gawler Craton and Curnamona Craton.

ma¢cs. The ¢nal voluminous magmatic event of the eastern and northern Gawler Craton is the anorogenic V1600^1575 Ma Hiltaba Suite and associated Gawler Range Volcanics (Fig. 3). The Adelaide Fold Belt contacts the eastern Gawler Craton and Stuart Shelf along the Torrens Hinge Zone, which marks the western limit of Neoproterozoic sedimentation in the Adelaide Geosyncline (see Fig. 3). Late Palaeoproterozoic^ Mesoproterozoic basement inliers within the fold belt include the northern Mount Painter Province ; a Palaeoproterozoic metasedimentary sequence intruded by 1575^1555 Ma granites and volcanics. In the northeast, the Willyama Inliers (which include the Olary and Broken Hill Blocks) expose Palaeoproterozoic sedimentary and volcanic units intruded by granites at V1720^ 1700 Ma and V1590 Ma [21]. These basement terranes are unconformably overlain by a thick Neoproterozoic to Early Cambrian cover sequence of clastics, carbonates and minor volcanics deposited within an aborted, intracratonic rift forming the Adelaide Geosyncline. Both basement

EPSL 5622 3-11-00

N. Neumann et al. / Earth and Planetary Science Letters 183 (2000) 107^120

and cover were deformed and partly metamorphosed during the Delamerian Orogeny (V500 Ma) to form the Adelaide Fold Belt. The Curnamona Craton comprises eastern^central South Australia and is bounded by the northern Adelaide Fold Belt to the west and the Willyama Inliers to the south (Fig. 3). Although poorly exposed, the Curnamona Craton appears to be of similar age to the eastern Gawler Craton and Willyama Inliers. 4. Surface heat £ow in South Australia Existing heat £ow values from 22 di¡erent lo-

111

cations within South Australia (Table 1) increase from west to east, with a distinct rise in values evident between the western and eastern Gawler Craton (Fig. 4). The zone of elevated heat £ow, which we term the South Australian heat £ow anomaly (SAHFA), overlaps the boundary between the eastern Gawler Craton and the Adelaide Fold Belt, including the Stuart Shelf (Fig. 3). Heat £ow measurements at the Olympic Dam Cu^U^Au-REE deposit on the Stuart Shelf provide the most detailed heat £ow dataset within the SAHFA. Measurements from seven locations excluding the Olympic Dam deposit yield a heat £ow average of 73 mW m32 , while the heat £ow recorded at the orebody is 125 mW m32 [23].

Table 1 South Australian heat £ow data Province

Site

Lat (³S)

Long (³E)

WGC

Maralinga Tarcoola Wudinna Iron Knob Whyalla Kadina Bute Bute Wokurna Olympic Dam RD21 Andamooka BLD1 WRD2a IDD1a ACD3a SGD2a HHD1a BD2a Parabarana Hill Ediacara Mount McTagga Carrieton Stockyard Kanmantoo Bendigo St Mootooroo Radium Hill Broken Hill Mount Gambier

30.17 30.62 32.98 32.72 33.17 33.97 33.87 33.93 33.72

131.60 134.50 135.55 137.13 137.50 137.75 138.02 137.97 138.12

29.98 30.60 30.45 32.55 34.77 35.08 33.20 32.25 32.50 31.95 37.75

139.72 138.12 139.30 138.48 138.80 139.25 139.47 140.93 140.50 141.47 140.86

EGC

SS

MPP AFB

WI SESA

n

Q (mW m32 )

20 17 6 2

4

12 11 4 56

54 49 58 109 91 101 88 87 91 125 67 78 62 101 84 86 50 126 96 101 92 88 88 64 68 75 80 92

H (WW m33 ) 2.7 4.9 7.5 8.4

7.9

3.8 3.1 3.8 4.6

Source [22] [22] [22] [22] [22] [22] [22] [22] [22] [23] [23] [23] [23] [23] [23] [23] [23] [22] [22] [22] [22] [22] [22] [22] [22] [24] [25] [22]

n = number of heat £ow determinations. Q = surface heat £ow. H = surface heat production. Provinces: WGC = western Gawler Craton, EGC = eastern Gawler Craton, SS = Stuart Shelf, MPP = Mount Painter Province, AFB = Adelaide Fold Belt, WI = Willyama Inliers, SESA = south^east South Australia. a Heat £ow measurements located within 3 km of Olympic Dam RD21.

EPSL 5622 3-11-00

112

N. Neumann et al. / Earth and Planetary Science Letters 183 (2000) 107^120

Fig. 4. Distribution of South Australian heat £ow records, with the main crustal element boundaries. See Fig. 3 for geological detail. Shaded area represents outcrop extent of Neoproterozoic sedimentation. Dashed lines as for Fig. 3. Province abbreviations: WGC = western Gawler Craton, EGC = eastern Gawler Craton, SS = Stuart Shelf, MPP = Mount Painter Province, AFB = Adelaide Fold Belt, WI = Willyama Inliers, CC = Curnamona Craton. Heat £ow data from [22^25].

Accompanying U^Th assay data suggest that V30 mW m32 of the additional heat £ow is generated within the orebody at depths less than 1 km, with the excess heat £ow presumably originating at greater depths [23,26]. The maximum heat £ow recorded within the SAHFA is 126 mW m32 from Parabarana Hill in the Mount Painter Province although, as outlined later, the thermal regime at this locality may be signi¢cantly modi¢ed by recent tectonism and hydrothermal activity. The eastern Adelaide Fold Belt, in the vicinity of the Willyama Inliers, is characterised by somewhat lower, but still elevated heat £ows of 68^81 mW m32 (Fig. 4). Regression of the South Australian heat £ow data for which basement heat production measurements are reported yields a poorly constrained slope (or characteristic length-scale, hr ) of 9.7 km, and intercept (or reduced heat £ow, qr ) of 29.5 mW m32 . The poor ¢t is probably due to a combination of factors, including the

incorporation of measurements from di¡erent heat £ow provinces (see below). The mean and standard deviation of all 22 heat £ow records from South Australia is 86 þ 20 mW m32 . In order to further de¢ne the character of the SAHFA, we will focus on heat £ow measurements within this region. Exclusion of the western Gawler Craton (n = 3, 54 þ 5 mW m32 ) increases the SAHFA to 91 þ 17 mW m32 (Fig. 5). By further excluding the somewhat lower values from the Willyama Inliers and eastern Adelaide Fold Belt (n = 4, 72 þ 7 mW m32 ) and the anomalously high measurements from Mount Painter, the Olympic Dam deposit and Mount Gambier, the remaining 11 records de¢ne a mean heat £ow of 92 þ 10 mW m32 (Fig. 5). This is our best estimate for the regional character of the SAHFA. The extent of the SAHFA is not precisely constrained from the available data (Figs. 2 and 4). It clearly extends from Adelaide in the south some 600 km north to the Mount Painter Province. North of Mount Painter it appears to merge with a broad heat £ow anomaly extending almost all the way across the continent (Fig. 2), although the density of data in the interior of the continent is too low to critically evaluate this. The most westerly record within the SAHFA is at Iron Knob, although the anomaly may well extend as far westwards as 136³E, which represents the ap-

Fig. 5. Surface heat £ow values as a function of longitude across South Australia. The three boxes denote the main regional heat £ow groups (province abbreviations as for Fig. 4), with the average heat £ow value for each region and number of measurements in brackets. Heat £ow data from [22^25].

EPSL 5622 3-11-00

N. Neumann et al. / Earth and Planetary Science Letters 183 (2000) 107^120

proximate longitude of the boundary between the eastern and western Gawler Craton (Fig. 4). Although the eastern boundary is relatively poorly de¢ned, the anomaly extends as far eastwards as the Mount Painter Province at V140³E giving a minimum width of 250 km. The available data suggest that the Willyama Inliers are characterised by somewhat elevated heat £ows, and thus could also be included in the SAHFA, in which case the anomaly is over 500 km across. Heat £ow values in the SAHFA are extraordinary in the context of averages reported for other Proterozoic terranes. For example, the Grenville province of North America is probably the best characterised Late Proterozoic province, with 30 heat £ow determinations providing an average of 41 þ 11 mW m32 [27] which makes it one of the `coldest' known continental regions. In view of the elevated heat £ow signature in South Australia, it is appropriate to assess whether these values are a true re£ection of the regional lithospheric thermal budget. In principle, the anomalous heat £ow could be the result of a number of factors including recent tectonism or magmatism, hydrologic processes, abnormal mantle heat £ow and/or exceptional concentrations of heat-producing elements in the crust. In this section, we discuss the evidence pertaining to these possibilities. 4.1. Recent tectonism The SAHFA overlaps a region of mild neotectonic activity which has resulted in the formation of the Flinders Ranges. It is broadly coincident with, and represents a reactivation of the Adelaide Fold Belt, which contrasts the adjacent Gawler Craton and Stuart Shelf, where there is no evidence of signi¢cant fault-related displacement since the Mesoproterozoic. The age of the initial uplift of the modern Flinders Ranges is still debated and may be as old as Eocene, or as recent as late Pliocene [28]. The coincidence of the uplands with a belt of active low-level, seismicity supports the notion that tectonism is ongoing, as do widespread Pleistocene fanglomerate aprons in the footwall to steep reverse faults that typically bound the uplands. While transient thermal anomalies associated with denudation of this up-

113

land system cannot explain the elevated heat £ow in the eastern Gawler Craton (Fig. 4), they may be responsible for some anomalies within the Flinders Ranges. In particular, the very high heat £ow (126 mW m32 ) from the Mount Painter region may be in£uenced by such recent tectonism, as this measurement is located on the bounding escarpment of the upland system. 4.2. Magmatic activity Southeastern Australia contains a well de¢ned, recent volcanic province that includes the very south^east corner of South Australia. [29]. While this magmatic province is also associated with elevated heat £ow (e.g. Mount Gambier [22]), the province is more than 400 km from the nearest measurement used to de¢ne the SAHFA and up to 1000 km from its north^westerly limit. This magmatism can therefore be dismissed as a significant contributor to the SAHFA. An older period of magmatism is represented by the Jurassic Wisanger Basalt at Kangaroo Island [30], however, this is again well south of any measurement used to de¢ne the SAHFA and is su¤ciently old that any associated transient thermal anomalies would have decayed by now. 4.3. Hydrologic activity The northeastern interior of South Australia contains Mesozoic sedimentary basins associated with the vast aquifer system of the Great Artesian Basin. Thermal gradients in these basins often exceed 50³C km31 and are locally as high as 70³C km31 [31]. These basins extend as far south as the northern limits of the Stuart Shelf and the northern Adelaide Fold Belt. However, none of the heat £ow measurements that de¢ne the SAHFA is from areas where these aquifers are known to exist. Moreover, the majority of the measurements are within crystalline basement, where there is no independent evidence for signi¢cant hydrologic activity. The only documented example of hydrothermal activity within the heat £ow anomaly is from Parabarana Hill, where the nearby active Paralana Hot Springs may signi¢cantly modify the measured heat £ow.

EPSL 5622 3-11-00

114

N. Neumann et al. / Earth and Planetary Science Letters 183 (2000) 107^120

4.4. Abnormal mantle heat £ow A number of observations and inferences bear on the thermal regime of the upper mantle beneath the SAHFA. Firstly, the absence of signi¢cant tectonism and magmatic activity in the region since the early Phanerozoic suggests that the mantle lithosphere is in no way thermally anomalous. By analogy with provinces of similar age elsewhere in the world, we might expect relatively thick lithosphere at least beneath the eastern Gawler Craton, where the last tectonic activity of any signi¢cance was in the Mesoproterozoic. This notion is corroborated by estimates of upper mantle seismic velocity [32] which show that the region is underlain by material of relatively fast seismic velocity at depths of 100^ 200 km. This contrasts with the more easterly parts of Australia, which are characterised by signi¢cantly slower upper mantle velocities. The boundary between these various mantle domains coincides with the eastern limit of Proterozoic crust, some 500 km east of the SAHFA. The relatively fast upper mantle velocities beneath the SAHFA suggest relatively cool upper mantle temperatures, and can be used to constrain the contributions of mantle and crustal heat sources to the observed surface heat £ow. Here we outline some simple calculations that allow estimation of the contributions assuming that, to account for the fast seismic velocities, the uppermost mantle must be no hotter than V500³C. We begin by noting that for a given surface heat £ow (qs ) and conductivity (k), Moho temperatures depend on both the heat £ow at depth (qm ), and the way in which the crustal contribution to heat £ow, qc (where qc = qs 3qm ) is distributed within the crust as represented by the length-scale, h [33]. For the thermal arguments developed below this length parameter is given by the mean depth de¢ned as: 1 hˆ qc

Zzc …H…z†z†dz 0

where H is the heat production rate and zc is the base of the crust, beneath which there is assumed

to be negligible heat production. Sandiford and McLaren [33] show that with h de¢ned in this way, the temperature at any depth, z, beneath the heat-producing parts of the crust is given by: T…z† ˆ …qm z ‡ qc h†=k For any assumed functional heat production distribution, the parameter h can be easily related to parameter hr inferred from regression of surface heat production and heat £ow data [33]. For the exponential model of Birch et al. [34] and others, h = hr . For a heat production distribution which consists of a layer of homogeneous heat production of thickness hr , h = hr /2. We now consider the thermal consequences at upper mantle depths (V40 km) of the various combinations of qm , qc and h consistent with the SAHFA where qs is V90 mW m32 . h can be constrained by recognising that the crustal contribution to the surface heat £ow qc = Hs h, where Hs is the characteristic surface heat production, and assuming the functional form of the heat production distribution. Regression of the surface heat £ow^heat production data suggests the SAHFA is characterised by hV10 km and qr V30 mW m32 , with a resulting qc of 60 mW m32 . For these parameters, T40 km is estimated to be 600³C for an assumed exponential distribution or 500³C for an assumed homogeneous distribution (Fig. 6). While these calculations are sensitive to the assumed thermal conductivity, about which there is some uncertainty, they suggest that for the relatively fast upper mantle seismic velocities, qm can be no more than 30 mW m32 . If the distribution more closely approximates an exponential distribution than a homogeneous distribution, then qm must be no more than about 15 mW m32 (Fig. 6). Note that if the elevated heat £ow was entirely due to anomalous mantle contributions (i.e. qm V60 mW m32 ), T40 km must exceed 800³C, which would seem to require signi¢cantly slower upper mantle velocities than suggested by Zielhuis and van der Hilst [32]. 4.5. Enrichment in crustal heat production The SAHFA forms part of a province that

EPSL 5622 3-11-00

N. Neumann et al. / Earth and Planetary Science Letters 183 (2000) 107^120

Fig. 6. Illustration of the e¡ect of partitioning of heat £ow between mantle and crustal sources on upper mantle temperatures. Figure indicates estimated temperatures at 40 km depth with a surface heat £ow of 90 mW m32 and a crustal heat production of 6 WW m33 , as suggested by Fig. 7. The heat production distribution, h, ranges from the exponential model (h = hr ) to the homogeneous model (h = hr /2). Assuming that these heat production distributions bracket the range in h that characterise the continental crust, the shaded area shows the range of Moho temperatures that could apply to the SAHFA. Regression of surface heat production^heat £ow data suggests qr = 30 mW m32 and h = 10 km. Equating qr with qm for these conditions yields Moho temperatures of 500^600³C depending on the distribution. If the surface heat £ow is entirely due to anomalous mantle heat contributions (i.e. qm = 60 mW m32 and qc = 30 mW m32 ), then Moho temperatures would exceed 800³C, while mantle heat £ows of 15 mW m32 (qc = 75 mW m32 ) yield Moho temperatures in the range 360^525³C.

hosts several signi¢cant uranium deposits, the most notable being the giant Olympic Dam deposit. Such occurrences suggest that the province may be unusually enriched in heat-producing elements, and it is therefore pertinent to explore the possibility that elevated crustal heat production provides the additional heat £ow within the SAHFA. If the anomalous heat £ow was entirely of a crustal origin, then these sources would be required to contribute at least 60 mW m32 . Assuming a length-scale, h, of 10 km then the characteristic surface heat production required to produce the anomaly would be 6 WW m33 , which is more than three times higher than the upper crustal heat production average of 1.8 WW m33 [5]. In order to evaluate this possibility, we have

115

compiled all existing geochemical data for Archaean to Mesoproterozoic lithologies from South Australia. Heat production values have been calculated as present day averages from U, Th and K2 O geochemical analyses (data from [35^39]; unpublished data). U, Th and K concentrations were obtained from whole rock powders using a X-ray £uorescence spectrometer. Detection limits for U and Th are 1.5 ppm, with an accuracy of þ 5% at 100U detection limit for U and Th and þ 0.02% for K2 O wt%. These detection limits and resultant uncertainties are only signi¢cant for samples with low U, Th and K concentrations. For instance, for U = 5 ppm, Th = 20 ppm and K2 O = 3.00 wt%, the calculated H is 3.0 WW m33 with an error of þ 0.5 WW m33 (17%). The error becomes more insigni¢cant at higher concentrations; for H = 6 WW m33 the error is only þ 8% assuming standard crustal ratios of U, Th and K2 O. For reference, average heat production values for typical lithologies range from 0.1 WW

Fig. 7. Surface heat production values from South Australian Precambrian basement terranes plotted as a function of longitude. Heat production values calculated from whole rock geochemical analyses. The upper bound of the shaded region is de¢ned by the binned average of the heat production data, the lower bound by the binned median. Note logarithmic scale for heat production values. The absence of data in the interval 138^139.5³E is due to the fact that in this region the Precambrian basement is covered by the Neoproterozoic sediments of the AFB. Province abbreviations as for Fig. 4. Data from [35^39]; unpublished data.

EPSL 5622 3-11-00

116

N. Neumann et al. / Earth and Planetary Science Letters 183 (2000) 107^120

Table 2 Average U, Th and K concentrations and heat production values for selected units of the SAHFA

EGC

MPP

K2 O (wt%)

Th/U

H (WW m33 )

25 27 26 39 42 17 61 22 93 46 41 85 61 55 18 7 74

1.78 4.48 5.23 5.02 3.49 3.92 5.16 4.83 5.85 5.38 5.32 5.47 4.47 5.64 3.85 1.97 3.57

3.6 5.4 4.3 3.5 4.7 2.4 6.1 2.2 2.5 5.1 4.1 5.0 3.6 3.9 1.3 3.5 3.5

3.9 3.7 4.0 6.2 5.8 3.4 7.6 4.8 17.0 6.1 6.0 11.1 9.4 8.3 5.3 1.2 11.3

288 131 423 33 58 35

4.17 6.37 5.66 5.78 5.24 6.50

3.8 3.0 3.6 5.5 5.8 2.7

41.0 21.6 61.6 4.5 7.2 6.7

Lithological groups

n

U (ppm)

Th (ppm)

Hutchison Group Donington Suite Myola Volcanics Moonta Porphyry McGregor Volcanics Wandearah Metasiltstone Middlecamp Granite Moody Suite Burkitt Granite Carappee Granite Hiltaba Suite-Coulta/Cleve Hiltaba Suite-Lincoln Hiltaba Suite-Moonta Hiltaba Suite-Stuart Shelf Basal Metasediments Freeling Heights Quartzite Mount Neill Granite, Pepegoona Porphyry Hot Springs Gneiss Box Bore Granite, microgranites Yerila Granite Wattleowie Granite Terrapinna Granite Petermorra Volcanics

13 40 3 20 10 13 25 27 11 2 15 16 22 54 18 3 60

7 5 6 11 9 7 10 10 37 9 10 17 17 14 14 2 21

42 30 28 27 29 7

75 44 116 6 10 13

n = number of samples used to calculate average concentrations of U, Th and K2 O. H = average current heat production. Units divided between the EGC and MPP (abbreviations as for Table 1) and listed in geochronological order. Data from [35^39] and unpublished data.

m33 for tholeiitic basalts and 2.1 WW m33 for shales, to 2.5 WW m33 for granites [40,41]. The surface heat production distribution of V2650 samples as a function of longitude clearly indicates that the SAHFA is characterised by elevated concentrations of heat-producing elements (Fig. 7). There is a distinct rise in median heat production from V3 WW m33 in the western Gawler Craton to V5 WW m33 for the eastern Gawler Craton. Surface heat production reaches a maximum within the Mount Painter Province where the range is extreme (1^65 WW m33 ), the average is V11 WW m33 and median is V6 WW m33 . To the east, heat production decreases to a median value of V2.5 WW m33 in the eastern Willyama Inliers. An obvious query about the treatment of the data using binned averages or means is that no account is taken of the relative volumetric contri-

bution of each of the rock types that comprise the various provinces. For much of South Australia poor outcrop precludes detailed assessment of area-integrated heat production rates. This is particularly true of the Gawler Craton, which is mostly covered by a thin veneer of aeolian sand. In contrast, excellent outcrop in the Mount Painter Province allows for the construction of relatively well constrained heat production maps. Using both airborne radiometric data and whole rock geochemistry, Sandiford et al. [42] were able to show that the area-integrated surface heat production for the province is 9.9 WW m33 . We suggest that the area-integrated heat production for the transect is bracketed by the binned mean and median heat production estimates, with representative upper crustal heat production falling in the region highlighted by shading in Fig. 7. In order to assess the extent to which the heat

EPSL 5622 3-11-00

N. Neumann et al. / Earth and Planetary Science Letters 183 (2000) 107^120

117

tion between observed heat £ow and heat production. Within the limits of the approach, this corroborates the notion that the South Australian heat £ow ¢eld is largely explicable in terms of crustal heat production distributions. While this applies to both the western Gawler Craton and SAHFA, it would appear that no single set of h^qm parameters can account for the heat £ow of both regions. In particular, the data seem to suggest that the western Gawler Craton is characterised by a slightly lower value of qm and/or h than the SAHFA. A lower value of qm would not be surprising given that the western Gawler Craton is believed to have formed in the Archaean and is therefore somewhat older than the provinces that comprise the SAHFA. In contrast with the western Gawler Craton and SAHFA, heat production values for the Willyama Inliers do not seem to be su¤cient to account for the heat £ow in this zone. Fig. 8. Surface heat £ow measurements plotted as a function of longitude. The shaded area is the calculated surface heat £ow obtained by using the binned average (upper bound) and median (lower bound) heat production data assuming in (a) a length-scale h = 10 km and mantle heat £ow qm of 30 mW m32 , and in (b) a length-scale h = 12.5 km and mantle heat £ow qm of 15 mW m32 (see text for discussion). Heat £ow data from [22^25].

production distribution may account for the observed heat £ow, we have modelled the surface heat £ow using the median and mean of the binned heat production data shown in Fig. 7, for two di¡erent sets of h^qm parameters. In Fig. 8a, we assume qm = 30 mW m32 and h = 10 km, while in Fig. 8b, we assume qm = 15 mW m32 and h = 12.5 km. In detail, the median heat production tends to slightly underestimate the heat £ow within the SAHFA (by V5^10 mW m32 ), slightly overestimate the heat £ow in the western Gawler Craton, and greatly underestimate the heat £ow in the Willyama Inliers (by up to 25 mW m32 ) for both sets of distribution parameters. In contrast, the mean heat production data tend to overestimate heat £ows in all but the Willyama Inliers. These ¢gures show that for all but the Willyama Inliers there is a ¢rst order correla-

5. Heat production in the SAHFA Given that the concentration of heat-producing elements in the SAHFA is anomalously high, it is appropriate to identify the source of the elevated heat production distribution. For this purpose, heat production values for the main Proterozoic lithologies in this region have been compiled (Table 2). Calculated heat production values for metasedimentary units within this transect are consistent with accepted lithological means and show little variation across provinces, suggesting that they do not contribute to the SAHFA. In contrast, granitic and volcanic lithologies display a large range in heat production, with many units within the eastern Gawler Craton and especially the Mount Painter Province containing extremely elevated values. Although U and Th concentrations within these granites are extreme, corresponding Th/U values are dominantly 3^5 (Table 2), suggesting that this enrichment is a primary magmatic signature. The heat production compilation highlights the fact that the region of high heat £ow in South Australia coincides with the occurrence of Proterozoic granites and volcanics which contain

EPSL 5622 3-11-00

118

N. Neumann et al. / Earth and Planetary Science Letters 183 (2000) 107^120

anomalously high U, Th and K concentrations. This suggests that Proterozoic felsic magmatic events have been important in shaping the vertical distribution of radiogenic elements in the crust. There is also good evidence for an increase in the upper range of heat production values for granites through time from the Archaean to the Mesoproterozoic, with the youngest granites in this region representing the ¢nal voluminous magmatic event in the Proterozoic. Within the SAHFA, heat production values for the Late Palaeoproterozoic Donington Suite average V4 WW m33 , increase to 5^8 WW m33 for V1700 Ma granites and remain high at V9 WW m33 for Hiltaba granites of the eastern Gawler Craton and Stuart Shelf (Table 2). As described in the Section 4, heat production reaches a maximum at Mount Painter, which also contains the youngest Proterozoic granites in the SAHFA. Because these heat production anomalies are geochemical phenomena and their source and distribution seem intimately associated with granite formation, the question of the origin of high heatproducing granites becomes vital. Although beyond the scope of this paper, we will make some observations here which are pertinent to the de¢nition of this problem. The geochemical characteristics of granitic magmas re£ect the composition of the source rocks and petrogenetic processes, such as fractional crystallisation, which modify elemental concentrations in the melt. Nd isotopic studies of granites are particularly useful in discriminating between mantle and crustal sources of di¡erent ages (e.g. [43]). The Nd isotopic record of Archaean^Mesoproterozoic granites from the SAHFA displays a large range in initial ONd from 2.8 to 311.1, suggesting considerable variations in the proportion and isotopic composition of crust and juvenile mantle material within these rocks [35^39,44]. Of particular interest are the initial ONd values of high heatproducing granites of the Mount Painter Province, which are quite high and show very limited variation from 33.2 to 30.5. Extreme fractionation of melts containing only a small enriched crustal component is required to reconcile high ONd values with elevated U, Th and K concentrations.

6. Discussion In the previous sections, we have shown that the high heat £ow of the SAHFA is largely the result of elevated crustal heat production, with inferences about the seismic velocity of the upper mantle supporting the notion that mantle heat £ow is less than 30 mW m32 . This suggests that the crust in this region contributes on average 60^75 mW m32 to the surface heat £ow, which is 2^3 times what would be expected on the basis of terranes of comparable or younger age in other continents. Such a high crustal radiogenic component implies that U, Th and K concentrations in the SAHFA are signi¢cantly higher than in typical crustal pro¢les, posing the question of how such elevated concentrations of heat-producing elements become localised in this crustal zone. Indeed, the concentrations amount to several times the crustal average and appear to require a source region of much greater volume than the directly underlying crust and mantle, especially given that lower crustal compositions are unlikely to represent a signi¢cant reservoir for radiogenic elements (e.g. [45]). The extraordinary enrichment in heat-producing elements evident in the SAHFA suggests that the heat production character of the continental crust is rather more variable than indicated in recent reviews of heat production distributions (e.g. [3,5]). While exceptional, the SAHFA does not appear to be unique. For example, Lewis and Hyndman [46] have documented heat £ow regimes from the northern Cordillera transect of Canada which are similar to the region described here. In this transect, heat £ow increases from 40^ 60 mW m32 in the Archaean North American Craton to V90 mW m32 across the Hottah and Fort Simpson terranes, reaching V120 mW m32 in the Nahanni terrane. Although these heat £ow values are consistent with elevated heat £ow measurements further west in the Cordillera, inferences on upper mantle thermal regimes derived from seismic velocity determinations suggest that the exceptional heat £ow in this region is due to elevated crustal heat production [46]. These occurrences suggest caution in the use of so-called `global averages', at least until the heat £ow data-

EPSL 5622 3-11-00

N. Neumann et al. / Earth and Planetary Science Letters 183 (2000) 107^120

set is extended to provide a more comprehensive coverage of regions outside North America and Europe. We have shown that elevated crustal heat production values need not lead to excessive upper mantle temperatures provided that mantle heat £ow is low and the heat production is mainly con¢ned to the upper crust. However, an important consequence of this enrichment is that thermal regimes at depth are extremely sensitive to small changes in the vertical distribution of crustal heat production, for example due to burial beneath a sedimentary pile. For the parameters appropriate to the SAHFA and thermal conductivities of 2.5^3 W m31 K31 , a change in heat production depth of 1 km produces a long-term increase in temperature of 20^30³C at any depth below the heat production. As outlined by Sandiford et al. [42], the sensitivity of crustal thermal regimes to the depth of the heat production distribution has important implications for the tectonic evolution of South Australia. By the earliest Phanerozoic, the high heat-producing Proterozoic basement of the SAHFA was di¡erentially buried by virtue of the distribution of sedimentation associated with the development of the Adelaide Geosyncline. During periods of subsequent tectonic compression (during the Delamerian Orogeny and more recently in the last 5 Ma), deformation has been strongly localised where this basement was most deeply buried. This re£ects a type of large scale basin inversion which is mechanically explicable in terms of the thermal consequences of di¡erential burial of high heat-producing basement terranes [42]. Acknowledgements We thank Primary Industries of South Australia (PIRSA) for access to the South Australia Geoscienti¢c GIS Dataset and Kathy Stewart for many useful discussions on the geochemistry of Proterozoic granites in South Australia. Hans Jurgen Fo«rster, Simon Turner and Ross Taylor are thanked for helpful reviews.[AH]

119

References [1] I. Vitorello, H.N. Pollack, On the variation of continental heat £ow with age and the thermal evolution of the continents, J. Geophys. Res. 85 (1980) 983^995. [2] P. Morgan, The thermal structure and thermal evolution of the continental lithosphere, in: H.N. Pollack, V.R. Murthy (Eds.), Structure and Evolution of the Continental Lithosphere, Phys. Chem. Earth 15, 1984, pp. 107^193. [3] R.L. Rudnick, W.F. McDonough, R.J. O'Connell, Thermal structure, thickness and composition of continental lithosphere, Chem. Geol. 145 (1998) 395^411. [4] S.R. Taylor, S.M. McLennan, The Continental Crust: its Composition and Evolution, Blackwell Scienti¢c Publications, Oxford, 1985, 312 pp. [5] S.M. McLennan, S.R. Taylor, Heat £ow and the chemical composition of continental crust, J. Geol. 104 (1996) 369^ 377. [6] J.G. Sclater, C.J. Jaupart, D. Galson, The heat £ow through oceanic and continental crust and the heat loss of the earth, Rev. Geophys. Space Phys. 18 (1980) 269^ 311. [7] H.N. Pollack, S.J. Hurter, J.R. Johnson, Heat £ow from the earth's interior: Analysis of the global data set, Rev. Geophys. 31 (1993) 267^280. [8] A.A. Nyblade, H.N. Pollack, A global analysis of heat £ow from Precambrian terranes: implications for the thermal structure of Archaean and Proterozoic lithosphere, J. Geophys. Res. 98 (1993) 12207^12218. [9] M.T. McCulloch, Sm^Nd constraints on the evolution of Precambrian crust in the Australian continent, AGU Geodyn. Ser. 17 (1987) 115^130. [10] L.O. Nicolayson, R.J. Hart, H.N. Gale, The Vredefort radioelement pro¢le extended to supracrustal strata at Carletonville, with implications for continental heat £ow, J. Geophys. Res. 86 (1981) 10653^10661. [11] L.D. Ashwal, P. Morgan, S.A. Kelley, J.A. Percival, Heat production in an Archean crustal pro¢le and implications for heat £ow and mobilization of heat-producing elements, Earth Planet. Sci. Lett. 85 (1987) 439^450. [12] D.M. Fountain, M.H. Salisbury, K.P. Furlong, Heat production and thermal conductivity of rocks from the Pikwitonei^Sachigo continental cross section, central Manitoba: implications for the thermal structure of Archean crust, Can. J. Earth Sci. 24 (1987) 1583^1594. [13] C. Pinet, C. Jaupart, The vertical distribution of radiogenic heat production in the Precambrian crust of Norway and Sweden: geothermal implications, Geophys. Res. Lett. 14 (1987) 260^263. [14] R.A. Ketcham, Distribution of heat-producing elements in the upper and middle crust of southern and west central Arizona: evidence from core complexes, J. Geophys. Res. 101 (1996) 13611^13632. [15] P. Morgan, J.H. Sass, Thermal regime of the continental lithosphere, J. Geodyn. 1 (1984) 143^166.

EPSL 5622 3-11-00

120

N. Neumann et al. / Earth and Planetary Science Letters 183 (2000) 107^120

[16] D.S. Chapman, K.P. Furlong, Continental heat £ow^age relationships, EOS Trans. AGU 58 (1977) 1240. [17] S. Ballard, H.N. Pollack, Diversion of heat by Archean cratons: a model for southern Africa, Earth Planet. Sci. Lett. 85 (1987) 253^264. [18] J.H. Sass, A.H. Lachenbruch, Thermal regime of the Australian continental crust, in: M.W. McElhinny (Ed.), The Earth ^ its Origin, Structure and Evolution, Academic Press, London, 1979, pp. 301^351. [19] J.F. Drexel, W.V. Preiss, A.J. Parker, The Geology of South Australia, Vol. I, The Precambrian, Geol. Surv. South Aust. Bull. 54, 1993, 242 pp. [20] S.J. Daly, C.M. Fanning, M.C. Fairclough, Tectonic evolution and exploration potential of the Gawler Craton, South Australia, AGSO J. Aust. Geol. Geophys. 17 (1998) 145^168. [21] P.M. Ashley, N.D.J. Cook, C.M. Fanning, Geochemistry and age of metamorphosed felsic igneous rocks with Atype a¤nities in the Willyama Supergroup, Olary Block, South Australia, and implications for mineral exploration, Lithos 38 (1996) 167^184. [22] J.P. Cull, An appraisal of Australian heat £ow data, BMR J. Aust. Geol. Geophys. 7 (1982) 11^21. [23] G.A. Houseman, J.P. Cull, P.M. Muir, H.L. Paterson, Geothermal signatures and uranium ore deposits on the Stuart Shelf of South Australia, Geophysics 54 (1989) 158^170. [24] J.H. Sass, J.C. Jaeger, R.J. Munroe, Heat £ow and near surface radioactivity in the Australian continental crust, United States Geological Survey, open ¢le report 76-250, 1976. [25] L.E. Howard, J.H. Sass, Terrestrial heat £ow in Australia, J. Geophys. Res. 69 (1964) 1617^1626. [26] K. Gallagher, Some strategies for estimating present day heat £ow from exploration wells, with examples, Exp. Geophys. 21 (1990) 145^159. [27] C. Jaupart, J.C. Mareschal, The thermal structure and thickness of continental roots, Lithos 48 (1999) 93^114. [28] V. Tokarev, M. Sandiford, V. Gostin, Landscape evolution in the Mount Lofty Ranges: implications for regolith development, in: G. Taylor, C. Pain (Eds.), Regolith '98: New Approaches to an Old Continent, 1999, pp. 131^139. [29] B.W. Smith, J.R. Presscott, Thermoluminescence dating of Mount Schank, South Australia, Aust. J. Earth Sci. 34 (1987) 335^342. [30] J.F. Drexel, W.V. Preiss, The Geology of South Australia, Vol. II, The Phanerozoic, Geol. Surv. South Aust. Bull. 54, 1995, 347 pp. [31] J.P. Cull, D. Conley, Geothermal gradients and heat £ow in Australian sedimentary basins, BMR J. Aust. Geol. Geophys. 8 (1983) 329^337. [32] A. Zielhuis, R.D. van der Hilst, Upper-mantle shear velocity beneath eastern Australia from inversion of waveforms from Skippy portable arrays, Geophys. J. Int. 127 (1996) 1^16.

[33] M. Sandiford, S.N. McLaren, Thermo-mechanical controls on heat production distributions and the long-term evolution of the continents, in: M. Brown, T. Rushmer (Eds.), Evolution and Di¡erentiation of the Continental Crust (submitted). [34] F. Birch, R.F. Roy, E.R. Decker, Heat £ow and thermal history in New England and New York, in: E.A. Zen, W.S. White, J.B. Hadley, J.B. Thompson (Eds.), Studies of Appalacian Geology, Northern and Maritime, Interscience, New York, 1968, pp. 437^452. [35] South Australia Geoscienti¢c GIS Dataset, CD ROM, Mines and Energy South Australia, 1997. [36] R.A. Creaser, The geology and petrology of middle Proterozoic felsic magmatism of the Stuart Shelf, South Australia, Latrobe University Ph.D. thesis (unpublished), 1989. [37] K. Stewart, High temperature felsic magmatism and the role of mantle magmas in Proterozoic crustal growth, the Gawler Range Volcanic province, University of Adelaide Ph.D. thesis (unpublished), 1992. [38] P.M. Ashley, Sodic granitoids and felsic gneisses associated with uranium^thorium mineralisation, Crockers Well, South Australia, Min. Depos. 19 (1984) 23^30. [39] B.F. Schaefer, Insights into Proterozoic tectonics from the southern Eyre Peninsula, South Australia, University of Adelaide Ph.D. thesis (unpublished), 1998. [40] R. Haenel, L. Rybach, L. Stegena, Handbook of Terrestrial Heat-Flow Density Determination; with Guidelines and Recommendations of the International Heat Flow Commission, Kluwer academic publishers, Dordrecht, 1988, 486 pp. [41] C.M.R. Fowler, The Solid Earth: An Introduction to Global Geophysics, Cambridge University, Cambridge, 1990, 472 pp. [42] M. Sandiford, M. Hand, S.N. McLaren, High geothermal gradient metamorphism during thermal subsidence, Earth Planet. Sci. Lett. 163 (1998) 149^165. [43] P.J. Patchett, Isotopic studies of Proterozoic crustal growth and evolution, in: K. Condie (Ed.), Proterozoic Crustal Evolution, Elsevier, 1992, pp. 481^508. [44] S. Turner, J. Foden, M. Sandiford, D. Bruce, Sm^Nd evidence for the provenance of sediments from the Adelaide Fold Belt and southeastern Australia with implications for episodic crustal addition, Geochim. Cosmochim. Acta 57 (1993) 1837^1856. [45] R. Rudnick, D. Fountain, Nature and composition of the continental crust: A lower crustal perspective, Rev. Geophys. 33 (1995) 267^309. [46] T. Lewis, R. Hyndman, High heat £ow along the SNORCLE transect, in: F. Cook, P. Erdmer (Eds.), Slave-Northern Cordillera Lithospheric Evolution (SNORCLE) Transect and Cordilleran Tectonics Workshop Meeting, Lithoprobe Report No. 69, University of Calgary, 1999, pp. 83^85.

EPSL 5622 3-11-00